Optimizing Preventive Adjuvant Linac-Based (OPAL) Radiation: A Phase 2 Trial of Daily Partial Breast Irradiation

Abstract

Purpose:

Evidence supports use of partial-breast irradiation (PBI) in the management of early breast cancer, but the optimal dose-fractionation remains unsettled.

Methods and Materials:

We conducted a phase 2 clinical trial (OPAL trial) to evaluate a novel PBI dosing schedule of 35 Gy in 10 daily fractions. Patients with close (<2 mm) margins also received a boost of 9 Gy in 3 fractions. Eligible patients underwent margin-negative lumpectomy for ductal carcinoma in situ or estrogen receptor–positive invasive breast cancer, up to 3 cm, pTis-T2 N0. The primary outcome was any grade ≥2 toxic effect occurring from the start of radiation through 6 months of follow-up. Secondary outcomes included patient-reported cosmesis, breast pain, and functional status, measured using the Breast Cancer Treatment Outcomes Scale, and physician-reported cosmesis, measured using the Radiation Therapy and Oncology Group scale. The Cochran-Armitage trend test and multivariable mixed-effects longitudinal growth curve models compared outcomes for the OPAL study population with those for a control group of similar patients treated with whole-breast irradiation (WBI) plus boost.

Results:

All 149 patients enrolled on the OPAL trial received the prescribed dose, and 17.4% received boost. The median age was 64 years; 83.2% were White, and 73.8% were overweight or obese. With median follow-up of 2.0 years, 1 patient (0.7%) experienced in-breast recurrence. Prevalence of the primary toxicity outcome was 17.4% (26 of 149 patients) in the OPAL trial compared with 72.7% (128 of 176 patients) in the control WBI-plus-boost cohort (P < .001). In longitudinal multivariable analysis, treatment on the OPAL trial was associated with improved patient-reported cosmesis (P < .001), functional status (P = .004), breast pain (P = .004), and physician-reported cosmesis (P < .001).

Conclusions:

Treatment with daily PBI was associated with substantial reduction in early toxicity and improved patient- and physician-reported outcomes compared with WBI plus boost. Daily external-beam partial-breast irradiation with 13 or fewer fractions merits further prospective evaluation.

Introduction

Most patients opting for breast conserving surgery will receive adjuvant whole-breast irradiation (WBI) followed by a tumor bed boost to optimize local control and breast cancer–specific survival.¹ Although overall well-tolerated, WBI plus boost can result in acute toxic effects such as painful dermatitis and fatigue and late toxic effects such as suboptimal cosmetic outcome. For example, in a prospective trial of WBI plus boost conducted at our institution, 65% of patients experienced breast pain, 85% developed fatigue, and 24% suffered from poor or fair long-term cosmetic outcomes.^2,3

With the recognition that most local recurrences after breast-conserving surgery alone occur near the original tumor bed, partial-breast irradiation (PBI), which targets only the tumor bed with margin, has been advocated to retain optimal local control while reducing acute and late radiation-associated toxicity.^4,5 Myriad approaches to delivering PBI exist, including intraoperative radiation therapy (RT), brachytherapy, and external-beam RT. The latter represents the most widely adopted method in North America; however, the optimal dose and fractionation for external-beam PBI remain controversial. Much of the initial PBI efforts focused on accelerated PBI in the form of twice-daily treatments for 5 days to a total dose of 35 to 38.5 Gy to render adjuvant treatment more convenient while retaining efficacy, but long-term results from the RAPID trial suggested this twice-daily dose-fractionation for external-beam PBI resulted in considerable decrements in cosmetic outcome versus WBI with or without boost.⁴

We hypothesized that external-beam PBI delivered with daily, as opposed to twice daily, treatments would allow for more complete normal-tissue repair between fractions and thus yield improved toxicity and cosmesis compared with traditional twice-daily external-beam PBI or WBI plus boost. To test this hypothesis, we conducted a single-arm, phase 2 trial using a novel 10-fraction regimen of daily PBI to 35 Gy with selective boost for patients with close (<2 mm) margins. We compared outcomes for our study population to a control group of similar patients treated with WBI plus boost on a prior prospective trial conducted at our institution.

Methods and Materials

Patient cohort

Between 2017 and 2018, 149 patients were enrolled on the OPAL (Optimizing Preventive Adjuvant Linac-Based radiation) trial, a multi-institutional, single-arm, phase 2 trial treating the partial breast to 35 Gy in 10 daily fractions. Patients were enrolled at The University of Texas MD Anderson Cancer Center (including the main campus located in the Texas Medical Center and 4 suburban integrated academic satellite campuses), Community Health Network (Indianapolis, Indiana); Summit Medical Group (Florham Park, New Jersey), Presbyterian Healthcare Services (Albuquerque, New Mexico), MD Anderson Cancer Center at Cooper (Camden, New Jersey), OhioHealth (Columbus, Ohio), and Baptist MD Anderson Cancer Center (Jacksonville, Florida). All relevant institutional review boards approved the trial.

Women eligible for enrollment were aged 50 years or older with pathologically confirmed ductal carcinoma in situ (pTis Nx M0) or invasive breast cancer (pT1-T2 N0[i– or i+] M0) treated with breast-conserving surgery with final negative margins (defined as “no tumor on ink”). Primary tumors were required to be ≤3 cm in size and macroscopically unifocal. Invasive disease was required to be estrogen-receptor positive, defined as ≥10% expression. The operative tumor bed was required to be visualized on computed tomography (CT) simulation, localized to 1 region of the breast, and amenable to partial-breast irradiation. Exclusion criteria included receipt of neoadjuvant chemotherapy, concomitant active treatment for another malignant condition or lesion, history of prior breast cancer, bilateral breast cancer, prior overlapping irradiation, pregnancy, or history of lupus or scleroderma. Race and ethnicity were determined from self-report and classified as Asian non-Hispanic (hereafter, Asian), Black non-Hispanic (hereafter, Black), Hispanic of any race (hereafter, Hispanic), and White non-Hispanic (hereafter, White).

Radiation treatment

Patients were treated to a total dose of 35 Gy in 10 once-daily fractions with external-beam radiation PBI. A 9-Gy boost delivered in 3 daily fractions was recommended for patients with a ≤2-mm surgical margin. Three-dimensional conformal radiation therapy, intensity modulated radiation therapy (IMRT), or volumetric modulated arc therapy (VMAT) were allowed. After CT simulation, the tumor bed, composed of the operative seroma and surgical clips, was delineated (Figs. E1-E3). A clinical target volume (CTV) was generated by creating a 2-cm volumetric expansion around the tumor bed, excluding the pectoralis muscle, lung, chest wall, ribs, and any tissue within 5 mm of the skin surface. A planning target volume (PTV) was created using a 5-mm expansion of the CTV in patients treated with deep inspiration breath hold and in whom target alignment with daily kilovoltage imaging to radiopaque clips in the surgical tumor bed could be used. For patients who could not be treated with deep inspiration breath hold and daily target localization, an 8-mm PTV expansion was used. For all patients, a PTV_EVAL structure, defined as the PTV excluding the pectoralis muscle, lung, chest wall, ribs, and any tissue within 5 mm of the skin surface, was delineated for aid in plan evaluation. The boost could be delivered with either electrons (2 cm radial expansion from tumor bed with or without scar to block edge) or photons (1 cm boost CTV expansion on tumor bed). Dose homogeneity and conformality were optimized by using forward or inverse planning of fields or field segments. Plan quality requirements included 95% of the PTV_EVAL receiving 95% of the prescription dose (≥33.25 Gy) and 98% of the CTV receiving 98% of the prescription dose (≥33.3 Gy). The volume of ipsilateral lung receiving at least 20 Gy was limited to <3%, receiving at least 10 Gy was limited to <10%, and receiving at least 5 Gy was limited to <20%. The mean heart dose was limited to <1 Gy for right-sided cancers and <2 Gy for left-sided cancers. The maximum point dose was limited to ≤108% of the prescription dose. For patients receiving a boost, 100% of the target volume received 90% of the prescribed dose.

Patient- and physician-reported assessments

Outcomes were assessed at baseline (preradiation), 6 months, 1 year, and annually thereafter. The Breast Cancer Treatment Outcomes Scale (BCTOS) was used to evaluate patient-reported cosmetic outcome, breast pain, and functional status.⁶ The BCTOS ranges from 1 to 4, with 1 indicating no difference between the treated and untreated breast, 2 indicating a slight difference, 3 indicating a moderate difference, and 4 indicating a large difference. Physician-reported cosmetic outcome was assessed by the treating radiation oncologist using the Radiation Therapy Oncology Group scale (1, excellent; 2, good; 3, fair; and 4, poor).⁷ Acute and late toxic effects were assessed weekly during irradiation, 3 weeks after completing radiation, and at the aforementioned time points posttreatment by the treating physician using the National Cancer Institute’s Common Toxicity Criteria for Adverse Events, version 4.0 (CTCAE), and the Subjective, Objective, Management, Analytical (SOMA) scale.⁸

Statistical methods

The primary outcome was specified as any CTCAE grade ≥2 toxic effect occurring within 6 months of starting PBI, and the primary objective was to compare incidence of this outcome for patients on the OPAL study with a control cohort composed of similar patients treated with WBI plus boost. The control cohort was selected from a recently reported clinical trial comparing conventionally fractionated WBI (CF-WBI) plus boost with hypofractionated WBI (HF-WBI) plus boost that was conducted at our institution (n = 287).^2,3 The WBI-plus-boost control cohort was limited to patients from this prior trial who met the following criteria: age ≥50 years, ductal carcinoma in situ (DCIS) or estrogen receptor–positive T1-T2 N0 or N1mic invasive cancer up to 3 cm in size, and no receipt of neoadjuvant chemotherapy. The trial recommended that no more than 1 cm³ of radiated tissue receive greater than 108% of the prescription dose. The comparison cohort was composed of 176 patients, 95 of whom received CF-WBI and 81 HF-WBI. In both arms, a tumor bed boost was used.

Differences in baseline patient, clinical, and treatment characteristics between the OPAL trial and the WBI-plus-boost control cohort were analyzed using the χ² or Fisher exact test, as appropriate. Differences in the primary outcome measure between the OPAL trial cohort and WBI-plus-boost control cohort were evaluated using the Cochran-Armitage trend test. An exploratory analysis compared the primary outcome for the OPAL trial cohort to patients treated only with HF-WBI plus boost, considering prior data indicating that early toxicity is less among patients treated with HF-WBI plus boost compared with CF-WBI plus boost.² We updated the posterior probability of grade 2 or higher toxic effects, comparing patients in OPAL with those in the WBI-plus-boost control group using a Bayesian approach, for which we assumed a β distribution with an uninformative prior. The Student t test compared means of BCTOS and physician-reported cosmesis at each prespecified time point, including baseline. Multivariable mixed-effects longitudinal growth curve models evaluated patient-reported outcomes and physician-reported cosmesis over time, adjusting for relevant covariates.⁹ Each model included a linear time term and baseline (preradiation) outcome measure as covariates. We considered candidate baseline patient and clinical variables with a P value <.2 on univariate analysis in the multivariable models. We then used a likelihood ratio test to determine whether the variable would be retained in the final model. We retained the covariates in the final model if P < .1. We expressed the results as estimates and standard errors with P values to indicate statistical significance. Given our recent work showing the adverse association of a large bra cup size with long-term patient-reported cosmesis among patients treated with WBI plus boost,¹⁰ we decided a priori to test in an exploratory analysis the interaction of bra cup size (A-C vs D or higher) with treatment group for patient-reported cosmetic outcome, with a plan to create stratified models if the P value for the interaction was <.10.

Oncologic outcomes, including in-breast tumor recurrence, contralateral breast tumor recurrence, regional nodal recurrence-free survival, distant metastasis-free survival, and overall survival, were calculated using the Kaplan-Meier method. Owing to the small number of events, oncologic outcomes on the OPAL trial were not compared with the control WBI-plus-boost cohort. P ≤ .05 was considered statistically significant. Statistical analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC). Data were locked for analysis in April 2021.

Results

Demographics and treatment of the OPAL cohort

Baseline demographics for the 149 patients treated on the OPAL protocol are presented in Table 1. A total of 118 patients (79.2%) enrolled at 1 of 5 MD Anderson locations within the Houston, Texas, area, and the remainder (31 [20.8%]) enrolled at affiliated institutions outside Texas. The median age was 64 years (range, 50–88 years), and 1.3% of the patients identified as Asian, 6.0% as Black, 9.4% as Hispanic, and 83.2% as White. Body mass index indicated overweight in 34.9% of patients and obesity in 38.9%. A total of 20.1% of patients had DCIS, and 79.9% had invasive cancer. Margins were close in 18.1%. Lymphovascular space invasion was present in 5.0% of patients with invasive cancer. Regarding surgery, 24.8% had an oncoplastic lumpectomy, 10.7% underwent placement of a BioZorb, and 4.7% had a concomitant breast implant.

Table 1. Baseline demographics for patients treated on the OPAL trial.

Factor	Patients
Sociodemographic
Age, y
Median (range)	64 (50–88)
50–59	52 (34.9)
60–69	58 (38.9)
≥70	39 (26.2)
Race and ethnicity
Asian	2 (1.3)
Black	9 (6.0)
Hispanic	14 (9.4)
White	124 (83.2)
Baseline BMI, kg/m2
12.5–24.9	39 (26.2)
25.0–29.9	52 (34.9)
≥30	58 (38.9)
Bra cup size
A	10 (6.7)
B	27 (18.1)
C	53 (35.6)
D or higher	59 (39.6)
Tumor
Breast tumor location
Upper inner	28 (18.8)
Upper outer	70 (47.0)
Lower inner	9 (6.0)
Lower outer	16 (10.7)
Central	26 (17.4)
Pathologic T stage
Tis	30 (20.1)
T1a or T1b	64 (43.0)
T1c	47 (31.5)
T2	8 (5.4)
Tumor size, cm
0–2.0	135 (90.6)
2.1–3.0	14 (9.4)
Grade
Low	60 (40.0)
Intermediate	72 (48.0)
High	17 (11.3)
Progesterone receptor
Positive	133 (89.3)
Negative	16 (10.7)
HER2-neu*
Negative	112 (94.1)
Positive	7 (5.9)
Final margin status
≥2 mm (negative)	122 (81.9)
<2 mm (close)	27 (18.1)
Lymphovascular space invasion*
No	107 (89.9)
Yes	6 (5.0)
Unknown	6 (5.0)
Treatment
Oncoplastic surgery
No	112 (75.2)
Yes	37 (24.8)
BioZorb placed at surgery
No	133 (89.3)
Yes	16 (10.7)
Concomitant breast implant
No	142 (95.3)
Yes	7 (4.7)
Radiation technique
3D-CRT	118 (79.2)
IMRT/VMAT	31 (20.8)
Boost delivered
No	123 (82.6)
Yes	26 (17.4)
Chemotherapy
No	142 (95.3)
Yes	7 (4.7)
Hormone therapy at 6 mo
No	19 (12.7)
Yes	130 (87.3)

Data are presented as n (%) unless otherwise indicated.

Abbreviations: 3D-CRT = 3-dimensional conformal radiation therapy; BMI = body mass index; IMRT = intensity modulated radiation therapy; OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation; VMAT = volumetric arc therapy.

^*Data presented are limited to the 119 patients with invasive disease.

All patients received 35 Gy in 10 fractions to the partial breast as stipulated by the protocol. A total of 26 patients (17.4%), all of whom had close margins, received a boost as stipulated by the protocol. The boost was inappropriately omitted in 1 patient with close margins. No patient with widely negative margins (≥2 mm) received a boost. A total of 79.2% of patients were treated with 3D-conformal radiation, and the remainder were treated with intensity modulated techniques. At the 6-month follow-up visit, 91.4% of patients with invasive cancer and 70.4% of those with DCIS were taking hormone therapy.

Comparison of baseline variables: OPAL cohort versus control WBI-plus-boost cohort

There were few differences between the OPAL trial cohort (n = 149) and the control WBI-plus-boost cohort (n = 176). The OPAL trial cohort included fewer patients with pT2 tumors (5.4% vs 9.7%; P = .03) and fewer patients treated with chemotherapy (4.7% vs 18.8%; P < .001), and the control WBI-plus-boost cohort included more patients with N1mic disease (0% vs 17.3%; P < .001). However, there were no significant differences in postoperative, preradiation patient-reported cosmetic outcome, functional status, or breast pain by treatment group or in physician-reported cosmesis by treatment group (Table 2).

Table 2. Comparison of OPAL and WBI plus boost cohorts.

Factor	Patients
	OPAL (n = 149)	WBI-plus-boost control (n = 176)	P value
Sociodemographic and clinical
Age, y
50–59	52 (34.9)	71 (40.3)	.10
60–69	58 (38.9)	76 (43.2)
≥70	39 (26.2)	29 (16.5)
Race and ethnicity
Asian	2 (1.3)	2 (1.1)	.93
Black	9 (6.0)	12 (6.8)
Hispanic	14 (9.4)	20 (11.4)
White	124 (83.2)	142 (80.7)
Baseline BMI, kg/m2
12.5–24.9	39 (26.2)	44 (25.0)	.39
25.0–29.9	52 (34.9)	51 (29.0)
≥30	58 (38.9)	81 (46.0)
Bra cup size
A	10 (6.7)	11 (6.3)	.94
B	27 (18.1)	36 (20.5)
C	53 (35.6)	64 (36.4)
D or higher	59 (39.6)	65 (36.9)
Preradiation physician-reported cosmetic outcome
Excellent	98 (65.8)	95 (54.0)	.05
Good	47 (31.5)	67 (38.1)
Poor	3 (2.0)	13 (7.4)
Fair	1 (0.7)	1 (0.6)
Tumor factors
Breast tumor location
Upper inner	28 (18.8)	22 (12.5)	.41
Upper outer	70 (47.0)	93 (52.8)
Lower inner	9 (6.0)	11 (6.3)
Lower outer	16 (10.7)	25 (14.2)
Central	26 (17.4)	25 (14.2)
Tumor size, cm
0–2.0	135 (90.6)	150 (85.2)	.14
2.1–3.0	14 (9.4)	26 (14.8)
Pathologic T stage
Tis	30 (20.1)	37 (21.0)	.03
T1a or T1b	64 (43.0)	50 (28.4)
T1c	47 (31.5)	72 (40.9)
T2	8 (5.4)	17 (9.7)
Pathologic N stage*
N0	119 (100)	115 (82.7)	<.001
N1mic	0 (0)	24 (17.3)
Grade
Low	60 (40.3)	52 (29.5)	.12
Intermediate	72 (48.3)	98 (55.7)
High	17 (11.4)	26 (14.8)
Progesterone receptor
Positive	133 (89.3)	148 (84.1)	.17
Negative	16 (10.7)	25 (14.2)
Unknown	0	3 (1.7)
HER2-neu*
Positive	7 (5.9)	6 (4.3)	.10
Negative	112 (94.1)	128 (92.1)
Unknown	0	5 (3.6)
Final margin status
≥2 mm (negative)	122 (81.9)	153 (86.9)	.21
<2 mm (close)	27 (18.1)	23 (13.1)
Treatment factors
Chemotherapy
No	142 (95.3)	143 (81.3)	<.001
Yes	7 (4.7)	33 (18.8)
Preradiation patient-reported outcomes
BCTOS, mean (95% CI) (number of patients)
Cosmetic outcome	1.54 (1.47–1.61) (n = 149)	1.60 (1.53–1.67) (n = 174)	.20
Functional status	1.33 (1.25–1.41) (n = 149)	1.26 (1.19–1.33) (n = 174)	.20
Breast pain	1.74 (1.63–1.85) (n = 149)	1.65 (1.57–1.73) (n = 174)	.15

Data are presented as n (%) unless otherwise indicated. Bold typeface indicates statistically significant comparisons.

Abbreviations: BCTOS = Breast Cancer Treatment Outcomes Scale; BMI = body mass index; CI = confidence interval; OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation; WBI = whole-breast irradiation.

^*Data presented are limited to the 119 patients (OPAL) and 136 patients (WBI-plus-boost control cohort) with invasive disease.

Primary outcome

Incidence of the primary outcome (defined as grade ≥2 CTCAE toxic effects occurring at any time between the start of radiation and the 6-month follow-up visit) was 17.4% (26 of 149 patients) for the OPAL trial compared with 72.7% (128 of 176 patients) in the WBI-plus-boost control group (P < .001). The Bayesian estimate of the primary outcome for patients in the OPAL trial was 18% (90% credible interval, 13%−23%), with a posterior probability of 1.0 that incidence of the primary outcome on the OPAL trial was less than the WBI-plus-boost control cohort. Individual toxic effects contributing to the primary outcome are presented in Table 3 and show lower risks of fatigue, pruritus, breast pain, shoulder arthralgia, dermatitis, hyperpigmentation, and breast edema for patients treated on the OPAL trial (all P < .05). In exploratory analysis, prevalence of the primary outcome was still favorable when the comparison was limited to patients treated with HF-WBI plus boost: 17.4% (26 of 149) versus 61.7% (50 of 81) (P < .001).

Table 3. Maximal toxic effects from treatment start to 6-month follow-up.

Toxic effect	Patients, n (%)		P value*
	OPAL (n = 149)	WBI-plus-boost control (n = 176)
Primary outcome (any grade ≥2 toxic effect)
No	123 (82.6)	48 (27.3)	<.001
Yes	26 (17.4)	128 (72.7)
Fatigue
None	73 (49.0)	27 (15.3)	<.001
Grade 1	70 (47.0)	128 (72.7)
Grade 2	6 (4.0)	17 (9.7)
Grade 3	0	4 (2.3)
Pruritus
None	136 (91.3)	53 (30.1)	<.001
Grade 1	12 (8.1)	114 (64.8)
Grade 2	1 (0.7)	9 (5.1)
Breast pain
None	88 (59.1)	46 (26.1)	<.001
Grade 1	59 (39.6)	116 (65.9)
Grade 2	2 (1.3)	14 (8.0)
Shoulder arthralgia
None	141 (94.6)	153 (86.9)	.047
Grade 1	8 (5.4)	21 (11.9)
Grade 2	0	2 (1.1)
Dermatitis
None	38 (25.5)	4 (2.3)	<.001
Grade 1	105 (70.5)	75 (42.6)
Grade 2	6 (4.0)	96 (54.5)
Grade 3	0	1 (0.6)
Hyperpigmentation
None	64 (43.0)	31 (17.6)	<.001
Grade 1	85 (57.0)	105 (59.7)
Grade 2	0	40 (22.7)
Breast edema
None	135 (90.6)	101 (57.4)	<.001
Grade 1	13 (8.7)	67 (38.1)
Grade 2	1 (0.7)	8 (4.5)
Wound complication
None	148 (99.3)	174 (98.9)	.66
Grade 1	1 (0.7)	2 (1.1)
Skin ulcer
None	149 (100)	172 (97.7)	.08
Grade 1	0	2 (1.1)
Grade 2	0	2 (1.1)
Seroma
None	135 (90.6)	138 (78.4)	.01
Grade 1	12 (8.1)	34 (19.3)
Grade 2	2 (1.3)	4 (2.3)
Breast infection
None	142 (95.3)	174 (98.9)	.13
Grade 2	6 (4.0)	2 (1.1)
Grade 3	1 (0.7)	0
Upper extremity edema
None	147 (98.7)	169 (96)	.32
Grade 1	2 (1.3)	6 (3.4)
Grade 2	0	1 (0.6)
Skin induration
None	122 (81.9)	149 (84.7)	.71
Grade 1	25 (16.8)	24 (13.6)
Grade 2	2 (1.3)	3 (1.7)
Telangiectasia
None	147 (98.7)	172 (97.7)	.63
Grade 1	2 (1.3)	3 (1.7)
Grade 2	0	1 (0.6)
Soft tissue fibrosis
None	112 (75.2)	135 (76.7)	.80
Grade 1	34 (22.8)	39 (22.2)
Grade 2	3 (2.0)	2 (1.1)
Deep tissue fibrosis
None	137 (91.9)	149 (84.7)	.02
Grade 1	10 (6.7)	27 (15.3)
Grade 2	2 (1.3)	0
Cough
None	145 (97.3)	174 (98.9)	.30
Grade 1	4 (2.7)	2 (1.1)
Dyspnea
None	147 (98.7)	175 (99.4)	.47
Grade 1	2 (1.3)	1 (0.6)
Nipple areolar change
None	132 (88.6)	144 (81.8)	.09
Grade 1	16 (10.7)	25 (14.2)
Grade 2	1 (0.7)	7 (4)
Breast atrophy
None	113 (75.8)	116 (65.9)	.05
Grade 1	28 (18.8)	38 (21.6)
Grade 2	8 (5.4)	22 (12.5)
Any other toxicity
None	139 (93.3)	149 (84.7)	.03
Grade 1	9 (6.0)	23 (13.1)
Grade 2–3	1 (0.7)	4 (2.3)

Abbreviations: OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation; WBI = whole-breast irradiation.

Patient-reported outcomes

Longitudinal patient-reported cosmetic outcome, functional status, and breast pain for patients on the OPAL trial versus the WBI-plus-boost control cohort are graphed in Fig. 1A-C. Mixed-effects growth curve models, adjusted for baseline differences in each outcome, confirmed an association of treatment on OPAL with improved outcomes for each domain: cosmetic outcome (effect size of treatment with OPAL vs the WBI-plus-boost cohort, −0.29; P < .001), functional status (−0.11; P = .004), and breast pain (−0.15; P = .004). Compared with bra cup size of A or B, a bra cup size of D or higher was associated with worse patient-reported outcomes in all 3 domains (P < .05 for all). The effect of time was significant for patient-reported functional status and breast pain (P < .001 for both), suggesting improvement over time for these 2 outcomes, but was not significant for cosmetic outcome (P = .37) (Table 4).

Fig. 1.

Longitudinal patient-reported outcomes and physician-reported cosmesis comparing patients treated on the OPAL trial to similar patients treated with whole-breast irradiation plus boost on a prior clinical trial. (A-C) Patient-reported cosmetic outcome, functional status, and breast pain, respectively, assessed using the Breast Cancer Treatment Outcomes Scale, with a score of 1 indicating no difference between the treated breast and the contralateral breast that served as an internal control, a score of 2 indicating a small difference, 3 indicating a moderate difference, and 4 indicating a large difference. (D) Cosmetic outcome rated by the treating physician, with a score of 1 indicating excellent outcome, 2 indicating good outcome, 3 indicating fair outcome, and 4 indicating poor outcome, using the Radiation Therapy and Oncology Group scale.⁵ *P < .05 for comparison of outcomes in OPAL patients compared with the whole-breast irradiation plus boost control cohort for each time point indicated, using the Student t test. Error bars indicate 95% confidence intervals. Abbreviation: OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation.

Table 4. Mixed effects growth curve models for patient-reported outcomes assessed using the Breast Cancer Treatment Outcomes Scale.

	Cosmetic outcome (n = 320)			Functional status (n = 320)			Breast pain (n = 320)
	Estimate	SE	P value	Estimate	SE	P value	Estimate	SE	P value
Time	−0.01	0.01	.37	−0.04	0.01	<.001	−0.06	0.02	<.001
Baseline	0.32	0.04	<.001	0.34	0.04	<.001	0.28	0.04	<.001
Treatment
WBI-plus-boost control	Referent	-	-	-	-	-	-	-	-
OPAL	−0.29	0.04	<.001	−0.11	0.04	.004	−0.15	0.05	.004
Age, y
50–59	Referent	-	-	-	-	-	-	-	-
60–69	−0.05	0.04	.25	-	-	-	−0.08	0.06	.18
≥70	−0.14	0.06	.01	-	-	-	−0.15	0.07	.03
Race and ethnicity
White	Referent	-	-	-	-	-	-	-	-
Asian	−0.20	0.17	.25	-	-	-	-	-	-
Black	0.02	0.08	.84	-	-	-	-	-	-
Hispanic	−0.18	0.06	.004	-	-	-	-	-	-
Bra cup size
A or B	Referent
C	−0.01	0.05	.93	0.004	0.05	.93	0.13	0.07	.05
D or higher	0.11	0.05	.03	0.10	0.05	.03	0.24	0.06	<.001
Margin status
Negative	Referent	-	-	-	-	-	-	-	-
Close	0.09	0.05	.10	-	-	-	-	-	-
Tumor location
Upper pole	Referent	-	-	-	-	-	-	-	-
Lower pole	-	-	-	0.08	0.05	.08	-	-	-
Central	-	-	-	0.06	0.05	.22	-	-	-
Pathologic T stage
Tis	Referent	-	-	-	-	-	-	-	-
T1a/b	−0.05	0.06	.39	−0.06	0.05	.26	−0.19	0.07	.009
T1c	−0.04	0.06	.45	−0.04	0.05	.40	−0.17	0.07	.02
T2	−0.09	0.08	.30	−0.02	0.08	.84	−0.11	0.11	.34
Chemotherapy
No	Referent	-	-	-	-	-	-	-	-
Yes	0.07	0.06	.30	0.04	0.06	.47	0.21	0.09	.02

Abbreviations: SE = standard error; OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation; WBI = whole-breast irradiation.

For the patient-reported cosmetic outcome, interaction of bra cup size with treatment group (OPAL vs the WBI-plus-boost control cohort) was marginal (P = .06). Multivariable mixed-effects models stratified by bra cup size indicated that patients treated with the OPAL regimen experienced a better cosmetic outcome than the WBI-plus-boost control cohort, regardless of bra cup size. However, the cosmetic benefits of the OPAL regimen were nearly doubled for women with a large bra cup size (D or higher) compared with a smaller bra cup size (bra cup size A, B, or C: effect size, −0.23; P < .001; bra cup size of D or higher: effect size, −0.41; P < .001).

When the multivariable analysis was limited to patients treated on the OPAL trial (n = 149), treatment with a boost was associated with a worse patient-reported cosmetic outcome (boost, compared with no boost: effect size, 0.17; P = .005). Neither treatment technique (VMAT/IMRT vs 3D), use of a BioZorb, oncoplastic surgery, nor presence of a breast implant were retained in final models for any patient-reported outcomes in models limited to the OPAL trial.

A final exploratory analysis was conducted comparing patients treated on the OPAL trial with boost (n = 26) compared with the WBI-plus-boost control cohort (n = 176) to delineate if treatment with PBI plus boost conferred different outcomes than WBI plus boost. In this analysis, cosmetic outcome favored PBI plus boost numerically but not statistically (effect size, −0.132; P = .11), and similar findings were noted for functional status (effect size, −0.055; P = .44) and breast pain (effect size, −0.117; P = .25).

Physician-reported outcomes

At the 2-year follow-up visit, cosmetic outcomes were improved for the OPAL trial versus the WBI-plus-boost control cohort as follows: excellent: 66.2% versus 52.2%; good: 32.3% versus 39.0%; fair: 0.8% versus 8.2%; and poor: 0.8% versus 0.6% (P = .004) (Table 5, Fig. 1D). A multivariable mixed-effects growth curve model confirmed substantial benefit associated with treatment on the OPAL trial for physician-reported cosmesis (effect size, −0.52; P < .001). When the multivariable analysis was limited to patients treated on the OPAL trial (n = 149), treatment with a boost was associated with a worse physician-reported cosmetic outcome (boost, compared with no boost: effect size, 0.19; P = .03). Neither radiation technique (VMAT/IMRT vs 3D), use of a BioZorb, nor presence of a breast implant were retained in the final model for physician-reported cosmetic outcomes in the OPAL trial.

Table 5. Physician-reported outcomes at the 2-year follow-up visit.

Outcome	Patients, n (%)		P value*
	OPAL (n = 149)	WBI-plus-boost control (n = 159)
Physician-reported cosmesis
Excellent	86 (66.2)	83 (52.2)	.004
Good	42 (32.3)	62 (39.0)
Fair	1 (0.8)	13 (8.2)
Poor	1 (0.8)	1 (0.6)
Not reported	19	17
SOMA scale
Telangiectasia
None	119 (91.5)	148 (93.1)	.56
Grade 1	7 (5.4)	9 (5.7)
Grade 2	3 (2.3)	0
Grade 3	1 (0.8)	2 (1.3)
Fibrosis
None	100 (76.9)	98 (61.6)	<.001
Grade 1	27 (20.8)	40 (25.2)
Grade 2	3 (2.3)	21 (13.2)
NCI CTCAE v4.0 scale
Fatigue
None	120 (92.3)	133 (83.6)	.03
Grade 1	9 (6.9)	22 (13.8)
Grade 2	1 (0.8)	4 (2.5)
Hyperpigmentation
None	107 (82.3)	112 (70.4)	.01
Grade 1	23 (17.7)	44 (27.7)
Grade 2	0	3 (1.9)
Skin induration
None	120 (92.3)	150 (94.3)	.66
Grade 1	9 (6.9)	7 (4.4)
Grade 2	1 (0.8)	2 (1.3)
Dermatitis
None	125 (96.2)	145 (91.2)	.09
Grade 1	5 (3.8)	14 (8.8)
Telangiectasia
None	122 (93.8)	151 (95.0)	.69
Grade 1	7 (5.4)	7 (4.4)
Grade 2	1 (0.8)	1 (0.6)
Skin ulceration
None	129 (99.2)	159 (100)	.26
Grade 1	1 (0.8)	0
Wound complication, noninfectious
None	130 (100)	159 (100)	NA
Breast infection
None	130 (100)	159 (100)	NA
Wound infection
None	130 (100)	159 (100)	NA
Upper extremity edema
None	130 (100)	158 (99.4)	.37
Grade 1	0	1 (0.6)
Breast edema
None	130 (100)	145 (91.2)	<.001
Grade 1	0	12 (7.5)
Grade 2	0	2 (1.3)
Superficial soft-tissue fibrosis
None	107 (82.3)	110 (69.2)	.009
Grade 1	22 (16.9)	45 (28.3)
Grade 2	1 (0.8)	4 (2.5)
Deep connective tissue fibrosis
None	123 (94.6)	138 (86.8)	.02
Grade 1	7 (5.4)	19 (11.9)
Grade 2	0	2 (1.3)
Seroma
None	127 (97.7)	149 (93.7)	.10
Grade 1	3 (2.3)	10 (6.3)
Breast pain
None	110 (84.6)	134 (84.3)	.71
Grade 1	19 (14.6)	23 (14.5)
Grade 2	1 (0.8)	2 (1.3)
Cough
None	129 (99.2)	158 (99.4)	.88
Grade 1	1 (0.8)	1 (0.6)
Dyspnea
None	129 (99.2)	157 (98.7)	.68
Grade 1	1 (0.8)	2 (1.3)
Nipple areolar changes
None	123 (94.6)	132 (83)	.001
Grade 1	7 (5.4)	20 (12.6)
Grade 2	0	7 (4.4)
Breast atrophy
None	104 (80.0)	104 (65.4)	.003
Grade 1	20 (15.4)	36 (22.6)
Grade 2	6 (4.6)	18 (11.3)
Grade 3	0	1 (0.6)
Other
None	127 (97.7)	142 (89.3)	.005
Grade 1	3 (2.3)	14 (8.8)
Grade 2	0	3 (1.9)

Abbreviations: NCI CTCAE v4.0 = National Cancer Institute Common Toxicity Criteria for Adverse Events, version 4.0; OPAL = Optimizing Preventive Adjuvant Linac-Based Radiation; SOMA = Subjective, Objective, Management, Analytical; WBI = whole-breast irradiation.

^*P value from Cochran-Armitage trend test.

Physician-reported toxic effects at the 2-year follow-up visit favored the OPAL trial for the following toxic effects: SOMA fibrosis, fatigue, hyperpigmentation, breast edema, superficial soft-tissue fibrosis, deep connective tissue fibrosis, nipple areolar changes, and breast atrophy. There were no toxic effects that favored the WBI-plus-boost control cohort (Table 5).

Oncologic outcomes

With a median follow-up of 2.0 years (interquartile range, 1.9–2.2 years), 1 patient in the OPAL trial experienced an in-breast tumor recurrence at 2 years’ follow-up, for a 2-year in-breast recurrence-free survival rate of 99% (95% CI, 95%−100%). This recurrence occurred in a patient with high-grade DCIS that measured 1.8 cm and was resected with negative (≥2 mm) margins. The recurrence was infield, with core biopsy showing recurrent, pure DCIS. No patients experienced regional nodal recurrence or distant recurrence. There was 1 contralateral breast cancer and 1 death not attributed to breast cancer or treatment toxicity.

Discussion

This article reports the primary outcome analysis of the OPAL trial, a 149-patient, phase 2, single-arm study that evaluated a novel 10-fraction schedule of daily partial-breast irradiation with 3-fraction boost limited to patients with close margins. In comparison with similar patients treated with WBI plus boost, we found that patients treated on the OPAL trial experienced considerable reduction in CTCAE toxic effects during and after treatment, with the primary outcome (grade ≥2 toxic effects noted within 6 months of starting treatment) occurring in 17.4% of patients treated on the OPAL trial versus 72.7% of patients in the WBI-plus-boost control cohort. Furthermore, we found that the OPAL regimen benefits extended broadly across multiple domains of patient-reported outcomes, physician-reported cosmesis, and CTCAE toxic effects through 3 years of follow-up. This evidence contributes to a growing body of literature supporting the benefits of daily partial-breast irradiation.

Prior randomized trials suggest that daily PBI confers cosmetic and toxicity benefits compared with WBI, whereas twice-daily partial breast irradiation confers worse cosmesis than whole-breast irradiation. For example, the IMPORT Low trial randomized patients to daily PBI (40 Gy in 15 fractions) versus the same dose fractionation of WBI without a boost. At 5 years’ follow-up, no differences were noted in local control, and 47.7% of patients treated with WBI experienced change in breast appearance compared with 35.1% of patients treated with PBI.¹¹ Similarly, investigators from the University of Florence randomized patients to every-other-day PBI (30 Gy in 5 fractions with IMRT) or WBI (50 Gy in 25 fractions) plus boost (10 Gy in 5 fractions) and found no significant difference in 10-year local control but significant improvements in both patient- and physician-reported cosmetic outcome in the PBI arm.¹² In contrast, the RAPID trial randomized patients to twice-daily PBI (38.5 Gy in 10 fractions) or standard-dose WBI (42.5 Gy in 16 fractions or 50 Gy in 25 fractions) with or without boost (10 Gy in 4–5 fractions) and found that treatment with twice-daily PBI approximately doubled the risk of poor or fair breast cosmetic outcome with a median follow-up of 7 years.⁴ Cosmesis data from the similarly designed Radiation Therapy and Oncology Group 0413 have yet to be published, although overall toxicity rates in that trial were not different for twice-daily PBI versus WBI plus boost.⁵ An interesting randomized comparison of daily versus twice-daily PBI found that treatment with twice-daily PBI led to marked increases in late skin toxic effects, fibrosis, and poor or fair cosmesis.¹³ Collectively, this evidence is consistent with findings from the OPAL trial, suggesting that daily (or every other day) PBI dosing schedules offer the most promise for improving the toxicity profile of PBI.

Considering that a chief goal of breast-conserving therapy is to achieve a cosmetic outcome that optimizes patient satisfaction, patient-reported outcomes are a key yardstick whereby various local therapy strategies should be evaluated. In our prior work, we found that large breast size was a key risk factor resulting in inferior patient-reported cosmetic satisfaction and breast pain among patients receiving WBI plus boost.¹⁰ For this reason, we sought to evaluate whether PBI may be a more favorable radiation treatment strategy, particularly for women with large bra cup size. Although our findings are not definitive, our exploratory analysis suggested that the cosmetic benefits of PBI were nearly twice as great among women with a bra cup size of D or higher compared with the benefits of PBI among women with a bra cup size of A through C. These findings thus suggest a strong rationale for preferring evidence-based daily PBI schedules, particularly among women with large breast size, as such schedules may considerably improve long-term quality of life for these patients compared with standard WBI plus boost.

Other challenges frequent in the management of early breast cancer include selecting the optimal treatment modality, managing breast implants, and indications for tumor-bed boost. In our study, we did not find substantial differences in patient-reported outcomes or physician-reported cosmesis by treatment modality. Although treatment with VMAT does often result in improved conformality compared with 3-dimensional techniques, we were not able to confirm a clinical benefit from this dosimetric advantage, although our study was underpowered to exclude a benefit and future research is ongoing. With regard to breast implants, capsular contracture resulting in impaired cosmesis and pain is a common occurrence after WBI, with a recent report noting a 25.4% rate of new or worse contracture among 70 patients with breast implants who underwent whole-breast irradiation, with median follow-up of 1.9 years.¹⁴ Only 7 patients in the OPAL study had breast implants, and to date, we were unable to document any worse outcomes in these patients versus patients without breast implants. Although this sample size was too small to yield definitive conclusions, our data justify continued prospective study of daily PBI in patients with breast implants, and we are optimistic that such an approach is likely to yield better cosmetic outcomes in this patient population compared with standard WBI plus boost. Finally, whereas delivery of a tumor bed boost is common after WBI, to our knowledge, this is the first study to integrate a tumor-bed-boost concept among patients receiving PBI. Our data demonstrated feasibility of this approach but also demonstrated a negative association between receipt of boost and cosmetic outcome among patients treated with PBI. However, it is worth noting that the effect size of PBI plus boost (0.19) was still considerably less than the effect size of treatment with WBI plus boost compared with the OPAL regimen (0.52), suggesting that patients still benefit cosmetically from PBI plus boost compared with WBI plus boost. Integrating a boost into PBI regimens could offer the benefit of selective dose escalation in patients with risk factors for local failure, and in our opinion, this merits ongoing study.

Based on the positive early results from the OPAL trial, we have designed and are currently enrolling to the OPAL II trial, a phase 3, noninferiority trial that enrolls women aged 40 years or older with early breast cancer to receive daily PBI with a dose of 26 Gy in 5 fractions versus daily PBI with a dose of 40 Gy in 15 fractions. This study design thus pits the dosing schedule from the UK Fast Forward Trial (which evaluated WBI treatment)¹⁵ versus that of IMPORT Low,¹¹ and if successful, it will provide novel evidence for a 1-week daily PBI regimen. Our study also incorporates use of a boost for women in their 40s or who have close margins. Similarly, investigators from the Juravinski Cancer Center are evaluating 2 dosing schedules for daily PBI in the Once-a-Day Accelerated Partial Breast Irradiation (OPAR) trial: 30 Gy in 5 fractions versus 27.5 Gy in 5 fractions, both delivered on a daily basis over 1 week (NCT02637024). Data from these trials will inform the next generation of PBI, which hopefully will allow many patients to realize the quality-of-life and convenience benefits of a short treatment course to the partial breast.

Several limitations of this work deserve mention. First, all the patients in our control cohort received a tumor bed boost. It is likely that a control cohort treated with WBI alone without boost would have experienced a better toxicity profile than WBI plus boost and that WBI alone and the OPAL regimen may be similar with regard to toxicity and patient-reported outcomes. Second, data regarding long-term local control for this dose-fractionation are needed before it should be considered a standard of care. Third, given coalescing data in support of 5-fraction approaches for both whole- and partial-breast irradiation, it is likely that the 10-fraction partial-breast approach studied in this trial will be less attractive. Fourth, some patients treated with WBI plus boost in our control cohort received treatment with high tangents to the low axilla as well, which could theoretically further worsen outcomes and accentuate differences between the OPAL trial and the WBI-plus-boost control cohort. Finally, it is worth noting that ongoing improvements in surgical technique—for example, increasing use of oncoplastic closure—may account for some of the measured differences in cosmetic outcomes between patients treated on the OPAL trial and those in the WBI-plus-boost control cohort.

Conclusions

Data from this prospective, phase 2 trial demonstrate that patients treated with a 10- to 13-fraction regimen of daily PBI experienced considerable short- and intermediate-term benefits in improved quality of life and reduced toxicity compared with similar patients treated with WBI plus boost. Daily PBI may be a particularly attractive option for women at risk for poor cosmetic outcome owing to risk factors such as a large bra cup size or breast implants.